the “evil god” argument

William Lane Craig has recently responded to the “evil god” argument put forward by Stephen Law in their debate, and Law has himself responded to that response.

I thought that now might be a good time to share my own thoughts on the argument, and where my agreements/disagreements with these two thinkers lay.

First of all, here’s a summary of a basic evidential argument from evil (as given in Philosophy of Religion by C. Stephen Evans & R. Zachary Manis):

1. If God exists he does not allow any pointless evil.
2. Probably, there is pointless evil in the world.
3. Therefore, God probably does not exist.

Craig’s response to this argument is to challenge the justification for (2) by noting our cognitive limitations in relation to God’s, employing what is known as the “skeptical theist” defence. The proponent of skeptical theism points out that the supposed justification for thinking that (2) is true is simply that it appears that it is. Sure, the skeptical theist says, we may utterly fail to discern any justifiable reason for, say, the agonising death of a deer whose legs are crushed beneath a fallen tree, but why think that our inability to discern such a reason is evidence that there is no reason? This failing on our part would only be evidence against there actually being such a reason if it’s likely that, were there a reason, we would be able to spot it. But why would we be likely to know the reasons why an all-intelligent being allows something horrible to occur? God’s cognitive capabilities far exceed our own, and thus it would hardly be surprising if God has reasons that we would fail to discern.

I think this a perfectly reasonable response to the argument. But here’s where Stephen Law’s “evil god” argument kicks in. He thinks we can run a parallel argument:

1. If evil god exists he does not allow any pointless good.
2. Probably, there is pointless good in the world.
3. Therefore, evil god probably does not exist.

In this argument, a particular good is pointless if it does not contribute to a greater evil. After all, evil god is perfectly bad, and entirely hate-filled, so he wouldn’t allow anything good unless it contributed to a greater evil. But it seems like there are goods that don’t contribute to any greater evil. What about some particularly beautiful yet undiscovered species of flower? What about the excessive cuteness of children? Aren’t these evidence against evil god?

Law notes that the same theodicies open to the defender of God are open to evil god (one can have a reverse free will defence, or a soul-destroying theodicy etc). But also, Law claims that the tactics of skeptical theism are just as available to the evil god proponent too. What justifies us in thinking that, just because many goods seem ‘pointless’, then many goods are pointless? Evil god is omniscient and supremely intelligent, so we can say in regard to him too that we aren’t likely to know whether or not there are greater evils that justify the goods. The arguments against both God and evil god seem to be even, Law says. BUT, he claims, nobody takes skeptical theism seriously in regard to discounting all the evidence against evil god. Nobody thinks that, after pointing to the smile of a child, or a stunning sunset, that a defender of the belief in evil god would be rational in saying, “hold on, we can’t recklessly draw such an inference here – we simply aren’t likely to be able to discern evil god’s reasons for permitting all this gratuitous good.” But since there is parity between the evidential argument from evil and the evidential argument from goodness, if skeptical theism doesn’t bail evil god out, it doesn’t bail God out either.

Craig denies that the defender of evil god would be irrational to appeal to skeptical theism. He claims that one simply cannot confidently conclude what the character of the creator is like based on the evidence of evil and good in the world. Which of them is correct? Is skeptical theism in defence of evil god justified? Or is the existence of inscrutable goods evidence against evil god? I actually think they’re both right.

It is quite clear to me that it is not at all likely that we’d be able to discern all the underlying reasons for an omniscient good being or an omniscient evil being to run the world as he does. But that isn’t the whole story. The above isn’t the only form an evidential argument from evil/goodness can take. I think skeptical theism shows this form to be a weak one, but a stronger one can be made through a comparative case.

the comparative argument

In a comparative argument, the God hypothesis, or the evil god hypothesis, is pitted against another hypothesis and judged on its power to explain some data, in this case, the existence of certain quantities/kinds of evil/good. Following Paul Draper, we can call this hypothesis “the Hypothesis of Indifference”, though I will describe it in my own way as follows:

HI: the universe is not ordered or guided by any all-powerful intelligent being that has a primary preference for good things, or a primary preference for bad things, nor a primary preference for the well-being of creatures, or a primary preference for the suffering of creatures.

HI is consistent with naturalism, and also a deism whereby god’s character is morally indifferent, but it isn’t consistent with theism, nor belief in evil god. Now, what about the data which the hypotheses will be “fighting” over? Let’s call the following observation about the world ‘O’:

O: There is roughly the same amount of good in the world as there is evil, and there is roughly the same amount of inscrutable evil/good as there is “meaningful” evil/good.

The truth of O is no doubt contentious, but let’s assume for the sake of argument that it’s true. It certainly doesn’t seem obviously false. Note: “meaningful evil/good” should here be understood as evil which, through an agent’s intention or not, contributes to a good, or good which, through an agent’s intention or not, contributes to an evil.

We can ask then, whether O is more surprising on theism/the evil god hypothesis, than on HI. For simplicity, we’ll keep to a comparison between HI and theism (the God hypothesis) – seeing as there is parity between the two, what counts for one will count for the other anyway. A comparative argument from evil might run like this then:

1. O is probable given HI.
2. O is improbable given theism.
3. Therefore, HI is a better explanation of O than theism.

Is (1) defensible? Well seeing as on HI there is no preference for either a wholly good, or wholly evil universe, it figures that, just by luck, the distribution of both would be about even. And since, on HI, there is no intelligent agent intending to bring about good or pleasurable states of affairs, there will likely be instances where evil or suffering fail to contribute to goods or pleasure, but also instances where such contributions occur via luck or incidentally, thus balancing out. So O is not all that surprising given HI. But is O surprising given theism?

I’d have to say that we can’t be sure. Again the considerations of skeptical theism (I think rightly) come into play. Can we really say with confidence that God couldn’t have achieved the amount of good there is (or will be) in the world without the amount of evil there is? I’d have to say that we can’t be so confident. I think the probability of O on theism is inscrutable. Thus (2) is a little strong and needs to be replaced by the following:

(2*): The probability of O on theism is inscrutable.

Does the conclusion still follow from (1) and (2*)? Yes, I think so. Put it this way; if you knew that one horse had a high chance of winning the race, but as for another, you had no clue as to their odds, you’d be a fool to pick the one you didn’t know about, even if in reality that horse had an even higher chance of winning. So yes, HI is a better explanation of O than theism. O is thus evidence against theism.

retaining rationality

Is HI’s explanatory power over theism in regards to O a problem for the theist? Does it make her theism irrational? I don’t think so. Perhaps relative to O theism is improbable, but theism is probable all things considered. Perhaps the theist’s being evidentially challenged here is no more significant than the kind of evidential challenge found in the following scenario:

Say (1) there is a pet store that I know is nearing the end of its transition from being an all-purpose pet store, to one that sells only cats and things relating to proper keeping of cats: the only non-cat left for sale is a solitary dog. Say also (2), I have a friend who is very familiar with the contents of this store, and is heading out to buy a single pet from there tomorrow. And (3) this friend has told me that she is going to buy that solitary dog. Now relative to (1) and (2), (4): “my friend is going to buy a cat” is much more probable than (3). (1) and (2) count as evidence against (3) in favour of (4). But surely I am rational in still believing that my friend is buying the dog. After all, I know that she is! But if I didn’t know (3), I would be much more justified in believing (4) over (3).

Similarly, O might not be a problem for the theist if she has good grounds for her theism! But if she didn’t know that theism were true, or didn’t have good grounds to so believe, she would be more justified in believing HI over theism. And by virtue of the symmetry of the circumstances, the same would be true of the evil god hypothesis. The evil god hypothesis isn’t the best explanation of O, but for someone who has good grounds for believing in evil god, this fact need not make the belief irrational.

So both Craig and Law are right. Skeptical theism is a perfectly legitimate move, but, nonetheless evil is evidence against God, and goodness is evidence against evil god. The agnostic, atheist and evil god proponent are justified in believing that evil counts against the existence of God. The agnostic, atheist and theist are justified in believing that good counts against the existence of evil god. But both the theist and evil god-believer need not be deterred when their own beliefs encounter this evidential challenge.


17 thoughts on “the “evil god” argument

  1. Pingback: Bill Craig loses a debate! (and all sorts of goodies are revealed) | ApologiaPad

  2. ‘Craig’s response to this argument is to challenge the justification for (2) by noting our cognitive limitations in relation to God’s, employing what is known as the “skeptical theist” defence’

    Is that the William Lane Craig who explained to Bart Ehrman that as a theist, he is entitled to assess the probability of God raising Jesus from the dead as very high?

    Or is it the William Lane Craig who claims he cannot assess what his god will do in any situation?

  3. Clearly you are wondering whether, if it is acceptable to be “skeptical” in regards to God’s intentions for permitting evil, then consistency demands such skepticism in regards to any assessment of the probabilities of God doing this or that. And of course, if this were so, such consistency would have consequences for assessing how the “religious context” of Jesus’ life affects the probability of his resurrection from the dead.

    This is an important issue which I’ve been somewhat reflecting on, and I have my suspicions that the skeptical theist isn’t committed to a wholesale skepticism of this sort, but my thinking on the matter is pretty fuzzy at the moment.

    What I will say is that I’m not sure how William Lane Craig himself understands the relation between the religious context of Jesus’ life, and the probability of God raising him from the dead. For instance, in Reasonable Faith he expresses caution over how confidently one can assess the prior probability of God raising Jesus from the dead;

    “Although Bayes’ Theorem can be useful for calculating the probability of some hypothesis on a given body of evidence and while philosophers such as Richard Swinburne have argued for the resurrection hypothesis by Bayesian means, professional historians do not really avail themselves of Bayes’ Theorem in the justification of historical hypotheses. One reason is that the values assigned to some of the probabilities involved are little more than conjectures. In the case of Jesus’ resurrection the probability of Jesus’ resurrection on the background information Pr(R|B) depends, we have seen, on the probability that God would raise Jesus of Nazareth from the dead Pr(R|G), which is speculative.” (pg359, third edition).

    I don’t know in what context he made that remark to Ehrman, but he may not have meant it in a bayesian framework.

    Personally, I’m tempted to view the “religious context” aspect of arguments for the historicity of the resurrection as providing a sort of intelligent design argument. In other words, the idea is that the apparent resurrection, in light of the religious context of Jesus’ life, seems contrived so as to perform an intelligent, religious function, which points to an intelligence (and a powerful one) as the cause of this otherwise freak phenomenon.

  4. Yo Martin, What is your take on Glenn Peoples insights into the issue and the debate he has had or will have with Stephen Law on the Unbelievable radio show.

    • Good question. The discussion hasn’t aired yet, but I’ve read the exchange they’ve had on Glenn’s blog. I think I’ll probably blog about it when I hear the discussion itself.

  5. Not quite analogous but I couldn’t resist recasting your example about the petstore:

    1. Say there is a civilization that is transitioning from preferencing dogmatic sectarian worldviews as sources of ultimate truth, to being one that accepts all of them merely as expressions of human cultural creativity.

    2. Say also I know a smart intelligent blogger, familiar with many religious and philosophical opinions, who intends to settle on a worldview as a source of truth from among those available in the said civilization.

    3. This blogger has said in my hearing that they intend to accept the Bible’s worldview as divinely inspired and the sole reliable approach to understanding the cosmos.

    Given 1 and 2,

    4. “The blogger will recognise that all religious worldviews are expressions of human culture and not reliable means of exclusive truth” is more probable.

    But given 3, the blogger thinks they have a rational basis to opt for the biblical worldview. The sheer intention of the blogger to make a selection in spite of the evidential probabilities, overrides the otherwise natural deduction that intelligent people who consider the options will not prefer dogmatic sectarian propositions over those evaluated with the minimum of cognitive bias.

  6. (I tried leaving this as a nested reply above but it seems to have been eaten by the comment machine. If it eventually appears, apologies for the double-posting, and please delete.)

    Martin, I’m sorry you took my parodic rewrite as insult. (And even more sorry that Rolo Baez seems to be baying for a spat off the back of it.) I’m more from the Nathan-telling-King-David-the-tale-of-the-lamb school of storytelling. If there was any derogation involved, it was to put down the idea that volitional bias should be allowed to override evidential probabilities. I had your other recent post about objectivity and values in mind as well. The best we can do in trying to be objective in these matters is to work ruthlessly to eliminate our own biases. In that regard, the volitional stance of the thinker/believer is a paramount consideration.

    By way of example, and since this post was partly to do with William Lane Craig, consider Craig himself. A high school debate team nerd (according to his own brother’s decription) who got converted to Evangelical Christianity in his teens, by all accounts for psychological rather than philosophical reasons, but who has spent his academic career shoring up his biases. Perhaps worst of all, he even deploys the “inner witness of the Holy Spirit” argument to head off any evidential possibility of falsification.

    You’ll notice in my parable I called the blogger smart and intelligent. I think you clearly are, as indeed is Craig, but I’m flagging up how value-based prior commitments will override the reasonable likelihoods of ending up with a dog or cat. Your last subhead was “Retaining Rationality”; that’s all I’m after too.

    • Hi Tim, I’m sorry if I jumped the gun in concluding that you were being insulting, and I’m sorry that it’s taken so long for your post to show up. Because it contained multiple links it was flagged as spam, and I don’t check through my spam that often.

      I think the kind of evidential considerations brought up here are different to the sorts of concerns I raised in my post about value and reason. In the example I gave here about the pet store, there was no ‘volitional’ bias or any other sort of cognitive bias at play, yet the belief that my friend was to buy the solitary dog was clearly rational despite being evidentially challenged in some way. So whatever we may think of how values ought to be involved in reasoning, value isn’t the factor at work here.

      Now then, on this separate topic of value and bias, you say, “the best we can do in trying to be objective in these matters is to work ruthlessly to eliminate our own biases.” I actually somewhat follow the Less Wrong community blog on human rationality, so I’m reasonably familiar with cognitive bias and the approach towards it which effectively says “yes, objectivity is hard, but we need to try as hard as we can to get it.” I’ll be honest, that blog has been the most challenging thing to my faith in a while. I recognise that I am biased. But I suppose I’m not sure that I agree neutralism really is the ideal for human rationality for reasons I explained in my post on value and reason. As a modern, neutralism has some pull on me, but I think it deserves critical scrutiny. No doubt you’ll view this as just another irrational example of motivated reasoning, but hey, that’s only bad if you beg the question against an alternate view ;)

      One more comment, this time on the link you posted in regards to the “internal witness of the holy spirit.” I am a ‘reformed epistemologist’ so I think that something like Plantinga’s account of how Christian belief is warranted is correct; I think Christian belief can be properly basic, and I don’t think that link really explained and refuted this position. However I don’t agree with Craig that Christian belief can be indefeasible. If Craig really did say that he would continue to believe in Christianity despite seeing Jesus’ body in the tomb with his own eyes, then I’d definitely part ways with him on that. Perhaps his experience of Christian belief is powerful enough to afford that, but mine isn’t.

  7. Pingback: Links and News — 19-Nov-11

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s